In magneto-optical disk devices which use the magnetic ultra high resolution method, a magneto-optical disk is used which is provided with a recording layer and with a reproducing layer having in-plane magnetization. In this type of magneto-optical disk device, during reproducing, a light beam is projected onto the reproducing layer side of the magneto-optical disk. Then, part of the area of the reproducing layer within the light beam spot is heated to above a predetermined temperature, and the magnetization of this portion (the aperture) shifts from in-plane magnetization to a perpendicular magnetization conforming to that of the recording layer beneath the aperture, i.e., the magnetization of the recording layer is copied to the reproducing layer. In this way, with this type of magneto-optical disk device, by reproducing the magnetization of the aperture, recorded marks smaller in diameter than the light beam spot can be reproduced.
In magneto-optical disk devices using this magnetic ultra high resolution method, it is preferable if the power of the light beam during reproducing (the reproducing power) is always at an optimum level. However, there are cases in which the optimum level of the reproducing power fluctuates with changes in the ambient temperature at the time of reproducing. For this reason, even if the current for driving the structure which produces the light beam (the driving current) is held constant, there are cases in which the reproducing power deviates from the optimum level.
If reproducing power is much stronger than the optimum level, the aperture formed on the magneto-optical disk becomes too large. Consequently, output of reproducing signals from tracks adjacent to the track being reproduced (crosstalk) is increased, the proportion of noise signals included in the reproduced data increases, and reading errors are more likely to occur.
Again, if reproducing power is much weaker than the optimum level, the aperture becomes smaller than the recorded mark, and the reproducing signal output from the target track is reduced. Accordingly, reading errors are more likely to occur in this case as well.
In a recording and reproducing device disclosed in Japanese Unexamined Patent Publication No. 8-63817/1996 (U.S. Pat. No. 5,617,400), in order to control reproducing power, long marks and short marks formed on a magneto-optical disk are reproduced. These long and short marks are two types of recorded marks for reproducing power control of different mark lengths. In this device, reproducing power is controlled so as to bring close to a predetermined value a ratio of the quantities of the reproducing signals from these recorded marks. By this means, in this device, reproducing power is maintained at an optimum value, and the likelihood of reading errors is reduced.
FIG. 30 is an explanatory drawing showing the general structure of this device. In this device, a light beam is projected from a semiconductor laser 108 onto a magneto-optical disk 112. Then, reflected light from marks for reproducing power control, which include long and short marks, is converted into a reproducing signal by a photodiode 113, and this reproducing signal is sent to an A/D (Analog/Digital) converter 115 and to a clock producing circuit 114. By means of the PLL (Phase Locked Loop) control method, the clock producing circuit 104 produces a clock signal synchronized with the reproducing signal, and sends this clock signal to the A/D converter 115.
Then, in accordance with the clock signal, the A/D converter 115 converts the reproducing signal into digital signals, which are sent to an amplitude ratio detecting circuit 116. The amplitude ratio detecting circuit 116 extracts, from the digital signals inputted for each clock signal, only the digital signals corresponding to upper and lower peak points. Then the amplitude ratio detecting circuit 116, based on the extracted digital signals, calculates the values of these upper and lower peak points and finds amplitude values for the long and short marks. Then a ratio between these amplitudes (amplitude ratio) is calculated and sent to a differential amplifier 110. This amplitude ratio corresponds with the size of the aperture on the reproducing layer of the magneto-optical disk.
The differential amplifier 110 compares the amplitude ratio with a predetermined standard value, and sends the results of this comparison to a reproducing power control circuit 111. The reproducing power control circuit 111 then controls driving current supplied to the semiconductor laser 108 in such a way that feedback reduces the difference between the amplitude ratio and the standard value.
In this way, the driving current supplied to the semiconductor laser 108 is controlled in such a manner that the light beam is always projected onto the magneto-optical disk at optimum reproducing power.
However, with this recording and reproducing device, the amplitude values of the recorded marks for reproducing power control are calculated using the values of only one upper peak point and one lower peak point. For this reason, the amplitude ratio calculated from these amplitude values is not sufficiently accurate, and thus there is a large error in control of reproducing power in this recording and reproducing device.
Again, as a method of reducing reading error rate with data recorded at high density, the PRML (Partial Response Maximum Likelihood) demodulating method has been proposed. The PRML demodulating method is a demodulating method in which a reproducing signal undergoes partial response equalization, and then maximum likelihood decoding (ML decoding) using Viterbi decoding.
A reproducing device using this demodulating-method is disclosed, for example, in Japanese Unexamined Patent Publication No. 6-243598/1994. In this device, a reproducing signal from an optical disk is equalized into PR(1,2,1) characteristics, and decoded into the most likely data by means of Viterbi decoding. FIG. 31 is an explanatory drawing showing the general structure of this device.
In reproducing using this device, an optical head 121 reads data recorded in an optical disk 120, and outputs an analog signal corresponding to this data. Then an A/D (Analog/Digital) converter 123 converts the analog signal into digital signals. The digital signals outputted by the A/D converter 123 are sent to a PRML demodulating circuit 126.
The PRML demodulating circuit 126 includes a PR equalizer 124 and a Viterbi decoder 125. The digital signals are equalized into PR(1,2,1) characteristics by the PR equalizer 124, and then Viterbi decoded by the Viterbi decoder 125, which outputs binarized data.
The analog signal outputted by the optical head 121 is also sent to a clock extracting section 122. The clock extracting section 122 produces and outputs to the A/D converter 123 clock signals with a bit cycle synchronized with the analog signal. The A/D converter 123 converts the analog signal to digital signals in accordance with the timing of the clock signal.
However, drawbacks of this reproducing device include the following. Namely, in this reproducing device, the sampling timing which is preferable for data reproducing, which is determined by the combination of the modulation method of the data recorded in the optical disk 120 and the demodulation method used by the PRML demodulating circuit 126, may not conform to the sampling timing which is preferable for accurately detecting the quantity of the reproducing signal of the recorded marks for reproducing power control.
Consider reproducing power when using PR(1,2,1)ML demodulating in the PRML demodulating circuit 126 to decode data from an optical disk recorded, for example, by the (1,7)RLL (Run Length Limited) modulation method.
FIG. 32 is an explanatory drawing showing, for this structure, the timing of A/D conversion (sampling) suited to PR(1,2,1)ML demodulating for a reproducing signal consisting of a pattern of repeated shortest marks (mark length 2Tc). As shown in FIG. 32, with sampling suited to PR(1,2,1)ML demodulating, a point at the shoulder of the reproducing signal is sampled.
On the other hand, the mark length of the short marks used for reproducing power control is typically 2Tc. Further, when reproducing these short marks, it is preferable to sample the upper and lower peak points of the reproducing signal obtained. However, as shown in FIG. 32, in sampling with this structure, a point at the shoulder of a reproducing signal corresponding to recorded marks 2Tc in length is sampled. Accordingly, it is not preferable to use the reproducing signal quantity obtained by this sampling for reproducing power control. Accordingly, a drawback of this structure is that, if A/D conversion is performed with a timing suited to PR(1,2,1)ML demodulating, it is difficult to perform reproducing power control.
In this example, a combination of PR(1,2,1)ML demodulating and (1,7)RLL modulation was considered, but for data reproduced by other combinations, too, such as PR(1,1)ML demodulating and the EFM (Eight to Fourteen Modulation) modulation method, there is a sampling timing preferable for the combination used.
In this way, conventional structures have the problem that, when the optimum sampling timing for data reproducing (which is determined by the combination of the PRML demodulating method and the modulation method) does not conform with the optimum sampling timing for reproducing power control, accurate reproducing is difficult.